--- title: "Combining 'Traditional' and Text-Based Models to Board Game Ratings" author: Brendan Graham date: '2022-01-26' slug: boardgames categories: - tidy tuesday - tidymodels - data science tags: - tidy tuesday - tidymodels - data science subtitle: summary: 'This post looks at a past [TidyTuesday](https://github.com/rfordatascience/tidytuesday) data set about board game ratings. After looking at the data I attempt to predict avereage board game' featured: no image: caption: '' focal_point: '' preview_only: no projects: [] draft: false ---
Last week I tried out a text based model for the first time. This week I want to continue working with a text based model, but supplement the text data with other non-text predictors. The goal will be to use the board game category (text data) and other non-text data to predict the average board game rating.
ratings %>%
select(-c(url, thumbnail)) %>%
add_table()
skimr::skim_to_list(ratings)
Variable type: character
| skim_variable | n_missing | complete_rate | min | max | empty | n_unique | whitespace |
|---|---|---|---|---|---|---|---|
| name | 0 | 1 | 1 | 107 | 0 | 21432 | 0 |
| url | 0 | 1 | 16 | 68 | 0 | 21831 | 0 |
| thumbnail | 6 | 1 | 135 | 139 | 0 | 21816 | 0 |
Variable type: numeric
| skim_variable | n_missing | complete_rate | mean | sd | p0 | p25 | p50 | p75 | p100 | hist |
|---|---|---|---|---|---|---|---|---|---|---|
| num | 0 | 1 | 10915.00 | 6302.21 | 0.00 | 5457.50 | 10915.00 | 16372.50 | 21830.00 | ▇▇▇▇▇ |
| id | 0 | 1 | 118144.78 | 105369.55 | 1.00 | 12308.50 | 104994.00 | 207219.00 | 350992.00 | ▇▂▃▃▂ |
| year | 0 | 1 | 1987.44 | 193.51 | 0.00 | 2001.00 | 2011.00 | 2017.00 | 3500.00 | ▁▁▇▁▁ |
| rank | 0 | 1 | 10916.00 | 6302.21 | 1.00 | 5458.50 | 10916.00 | 16373.50 | 21831.00 | ▇▇▇▇▇ |
| average | 0 | 1 | 6.42 | 0.93 | 1.04 | 5.83 | 6.45 | 7.04 | 9.57 | ▁▁▅▇▁ |
| bayes_average | 0 | 1 | 5.68 | 0.36 | 0.00 | 5.51 | 5.54 | 5.67 | 8.51 | ▁▁▁▇▁ |
| users_rated | 0 | 1 | 866.96 | 3679.82 | 30.00 | 56.00 | 122.00 | 392.00 | 108975.00 | ▇▁▁▁▁ |
details %>%
add_table()